How to use pyttsx to play the Chinese?

[Near-Tam]

Can pyttsx play Chinese? If so, what should do that? Do I need to install the Chinese library? If so, how do I download and install?

[nickgermaine]

I have no idea. I've just taken over this project, and honestly, haven't had much time to familiarize myself with the code yet.

[spreadred]

Perhaps this is pertinent?

https://stackoverflow.com/questions/18715563/how-to-tts-chinese-using-pyttsx

[WickyLeo]

voices = ['com.apple.speech.synthesis.voice.mei-jia',
'com.apple.speech.synthesis.voice.sin-ji.premium',
'com.apple.speech.synthesis.voice.ting-ting'];
In Mac OS，there are all chinese voice，choose one your want.and then:
engine = pyttsx.init()
engine.setProperty('voice', random.choice(voices))
engine.say(u'你好')
engine.runAndWait()

It‘s OK.

[sincethenn]

I want to play chances in ubuntu, I found it actually always said "chinese letter letter letter ..." and so on.
The last 2 voices are:
<Voice id=Mandarin

          name=Mandarin
          languages=[b'\x05zh']
          gender=male
          age=None>
<Voice id=cantonese
          name=cantonese
          languages=[b'\x05zh-yue']
          gender=male
          age=None>

The script is

import pyttsx3

engine = pyttsx3.init()
voices = engine.getProperty('voices')
for voice in voices:
    print(voice)
engine.setProperty('voice', voices[-2].id)
engine.say("你好")
engine.runAndWait()

[tomtan0817]

I run pyttsx3 in Ubuntu 20.04.4 LTS，the voices[ ].id is Mandarin. But after I trace the code, I find even I set the setProperty as Mandarin, pyttsx3 still use "default" to be get in getProperty. So try to modify the /usr/lib/aarch64-linux-gnu/espeak-data/voices/default as below.

name default
language asia-zh
gender male

And modify the python code "engine.setProperty('voice', voices[-2].id)" to "engine.setProperty('voice', 'zh')", then it works!

I want to play chances in ubuntu, I found it actually always said "chinese letter letter letter ..." and so on. The last 2 voices are: <Voice id=Mandarin

          name=Mandarin
          languages=[b'\x05zh']
          gender=male
          age=None>
<Voice id=cantonese
          name=cantonese
          languages=[b'\x05zh-yue']
          gender=male
          age=None>

The script is

import pyttsx3

engine = pyttsx3.init()
voices = engine.getProperty('voices')
for voice in voices:
    print(voice)
engine.setProperty('voice', voices[-2].id)
engine.say("你好")
engine.runAndWait()

[jfthuong]

I have created a small function help find a voice in Chinese:

import pyttsx3
from pyttsx3.voice import Voice

def get_chinese_voice(engine: pyttsx3.engine.Engine) -> Voice:
    voices = engine.getProperty("voices")
    for voice in voices:
        if voice.languages and voice.languages[0] == "zh-CN":
            return voice
        if "Chinese" in voice.name or "Mandarin" in voice.name.title():
            return voice

    raise RuntimeError(f"No Chinese voice found among {voices}")

You can use as such:

tts_engine = pyttsx3.init()
chinese_voice = get_chinese_voice(tts_engine)
tts_engine.setProperty("voice", chinese_voice.id)